View metadata, citation and similar papers at core.ac.uk brought to you by CORE
provided by Elsevier - Publisher Connector
Food Chemistry 164 (2014) 347–354
Contents lists available at ScienceDirect
Food Chemistry
journal homepage: www.elsevier.com/locate/foodchem
Analytical Methods Polyphenol extraction optimisation from Ceylon gooseberry (Dovyalis hebecarpa) pulp ⇑ Vivian Caetano Bochi a, , Milene Teixeira Barcia a, Daniele Rodrigues a, Caroline Sefrin Speroni b, M. Monica Giusti c, Helena Teixeira Godoy a
a Department of Food Science, Faculty of Food Engineering, Campinas State University (UNICAMP), P.O. Box 6121, 13083-862 Campinas, SP, Brazil b Integrated Center for Laboratory Analysis Development (NIDAL), Department of Food Technology and Science, Center of Rural Sciences, Federal University of Santa Maria, 97105-900 Santa Maria, RS, Brazil c The Ohio State University, Department of Food Science and Technology, 2015 Fyffe Road, Columbus, OH 43210, United States
article info abstract
Article history: Originally from Asia, Dovyalis hebecarpa is a dark purple/red exotic berry now also produced in Brazil. Received 27 October 2013 However, no reports were found in the literature about phenolic extraction or characterisation of this Received in revised form 8 May 2014 berry. In this study we evaluate the extraction optimisation of anthocyanins and total phenolics in Accepted 11 May 2014 D. hebecarpa berries aiming at the development of a simple and mild analytical technique. Multivariate Available online 16 May 2014 analysis was used to optimise the extraction variables (ethanol:water:acetone solvent proportions, times, and acid concentrations) at different levels. Acetone/water (20/80 v/v) gave the highest anthocyanin Keywords: extraction yield, but pure water and different proportions of acetone/water or acetone/ethanol/water D. hebecarpa (with >50% of water) were also effective. Neither acid concentration nor time had a significant effect Response surface methodology Simplex lattice design on extraction efficiency allowing to fix the recommended parameters at the lowest values tested Anthocyanin (0.35% formic acid v/v, and 17.6 min). Under optimised conditions, extraction efficiencies were increased Phenolic by 31.5% and 11% for anthocyanin and total phenolics, respectively as compared to traditional methods Extraction optimisation that use more solvent and time. Thus, the optimised methodology increased yields being less hazardous and time consuming than traditional methods. Finally, freeze-dried D. hebecarpa showed high content of target phytochemicals (319 mg/100 g and 1421 mg/100 g of total anthocyanin and total phenolic content, respectively). Ó 2014 Elsevier Ltd. All rights reserved.
1. Introduction the mouth due to a velvety texture and could be considered co- product (Morton & Dowling, 1987). Recently, it is being cultivated Ceylon gooseberry or Ketembilla (Dovyalis hebecarpa, Salicaceae as an exotic fruit in the southwest regions of Brazil. High produc- family), originally from Sri Lanka and introduced into USA (Florida) tion is obtained from March to May, but it is cultivated until in 1920, is a small spherical fruit (0.5–1 inch) characterised by a August with satisfactory harvesting yields. deep purple-red colour, sour juice, and small seeds enclosed in A hybrid of D. hebecarpa with Dovyalis abyssinica was reported the pulp. Fruits can be consumed as jams, juice, or fresh after skin as a source of vitamin C (120.3 mg/100 g of fresh fruit) and with removal. Despite an attractive appearance, skins are unpleasant in good physical quality for market, showing an average of 75% pulp (Cavalcante & Martins, 2005). Phenolic composition of this hybrid revealed higher contents of anthocyanins in fruit peels and carote- Abbreviations: SLD, simplex lattice design; RSM, response surface methodology; noids in the pulp (De Rosso & Mercadante, 2007). Nevertheless, no TMA, total monomeric anthocyanin; TPC, total phenolic content; GAE, gallic acid reports were found about extraction and phytochemicals content equivalent; CGE, cyanidin-3-glucoside equivalent. in D. hebecarpa species. ⇑ Corresponding author. Address: Santa Maria Federal University (Universidade Federal de Santa Maria, UFSM), Rural Science Center (Centro de Ciências Rurais, Phenolic compounds are secondary metabolites in fruits and CCR), Food Technology and Science Department (Departamento de Tecnologia e vegetables acting as a defense barrier against microorganism, Ciência de Alimento, DTCA), 42 Building, 97119-900 Santa Maria, RS, Brazil. Tel.: insects, and UV radiation, or as attractants to promote pollination +55 55 3213 4890; fax: +55 19 3521 2153. and seed dispersal. Some phytochemicals, as anthocyanins and E-mail addresses: [email protected] (V.C. Bochi), [email protected] phenolic acids, are very important for food acceptance acting as (M.M. Giusti), [email protected] (H.T. Godoy).
http://dx.doi.org/10.1016/j.foodchem.2014.05.031 0308-8146/Ó 2014 Elsevier Ltd. All rights reserved. 348 V.C. Bochi et al. / Food Chemistry 164 (2014) 347–354 natural pigments or adding specific flavours, such as astringency Due to the increasing need for efficient, simple, and mild and acidity (Shahidi & Naczk, 2003; Lattanzio, Kroon, Quideau, & extraction procedures for anthocyanins and other phenolics to Treutter, 2009). allow its further evaluation and characterisation, this work was Potential health benefits of phenolic compounds in the human focused on optimising extraction methods for these types of diet have been extensively discussed in the literature and are still compounds from D. hebecarpa pulp. Within this study, multi- being investigated mainly linked to antioxidant activity and variate experimental designs were used to develop procedures degenerative disease prevention (Zafra-Stone et al., 2007). The bio- investigating extraction variables to improve knowledge in this logical activity of berry phytochemicals is believed to be a result of field. multiple mechanisms that initially were believed to be mainly linked to antioxidant effects. However, nowadays the questionable 2. Material and methods polyphenols bioavailability has given rise to the discussion of different mechanisms by which it could act in the onset and in the 2.1. Chemicals and equipment development of degenerative diseases. Some of them are their anti-inflammatory activity, their enzyme inhibition capacity, The phenol reagent Folin–Ciocalteu and gallic acid were or their potential modulation of the gut microbiota (Chiva- obtained from Sigma–Aldrich. A Büchi rotary evaporator and an Blanch & Visioli, 2012). UV-1600 spectrophotometer from Proanálise (São Paulo, Brazil) Anthocyanins are glycosylated forms of polyhydroxy and were used for concentration and quantification purposes. A food polymethoxy derivatives of 2-phenylbenzopyrylium, acylated or processor (Philips Mini Food Processor, model HR7625) Terroni not, which are found as red to purple natural pigments in fruit Freeze-dryer, model LS-3000E (São Carlos, Brazil) and a Qhimis and vegetables. Usually, extraction procedures are conducted using analytical grinder, model Q298A (São Paulo, Brazil) were used for acidified solvents due to the higher anthocyanin stability at low pH sample preparation. values (Giusti & Jing, 2008). However, degradation of native chemical structure due to acid hydrolysis or oxidation can occur if 2.2. Sample preparation the acid concentration is high or if the extraction time is long (Naczk & Shahidi, 2006; Revilla, Ryan, & Martín-Ortega, 1998). Ripened samples (12.5 °Brix) were obtained from Bragança Thus, extraction conditions should be studied to ensure high Paulista city (São Paulo, Brazil) in April of 2010 (see Supplementary extraction yields and minimal degradation in native chemical material for fruit images). After harvesting, fruits were manually structure of these pigments. Finally, results obtained under peeled while frozen ( 20 °C) to minimise enzymatic degradation optimised extraction conditions will better represent the real com- and decrease juice loss. Frozen pulps were crushed using a food position of the vegetable matrix being used for characterisation processor and placed into trays to re-freeze. Materials were then studies. freeze-dried until the pressure was reduced to stable values lower Aiming to develop efficient and fast new extraction techniques, than 22lHg. Since particle size is extremely relevant to extraction sonication and subcritical fluids have been tested with promising effectiveness, freeze-dried samples were grinded until a fine and results (Adil, Cetin, Yener, & Bayındırlı, 2007). However, these flu- visually homogenous powder was obtained. ids increase the extraction costs and require equipment that is not typically present in most laboratory facilities. Moreover, in some cases, traditional methods still give better yields. Previous litera- 2.3. Initial extraction procedure ture data revealed that solvent extraction showed higher results A water/acetone (25:75 v/v) mixture, acidified with 2% of formic than an optimised subcritical fluid method (CO2 plus ethanol) in apple and peach pomace for total phenolics and antioxidant activ- acid was the initial solvent tested (Mertz, Cheynier, Günata, & Brat, ity (Adil et al., 2007). Nevertheless, the presence of even relatively 2007). Freeze-dried powder samples (0.5 g) were mixed with the safer solvents (such as alcohol) in the composition of natural colo- extraction solvent (15 ml) in a ratio of 1:30 w/v for 15 min, filtered rants may limit its application in food. Thus, even that solvent and re-extracted under the same conditions. Both filtrates were extraction generates reliable results and satisfactory yields it combined, taken to a volume of 100 ml in a volumetric flask, and should be optimised aiming the reduction or depletion of organic analysed for total phenolic and monomeric anthocyanins using dissolvent agents. In addition, most extraction procedures were the methods described below. not optimised or validated and few details are reported about the selection of extraction variables. Considering it, anthocyanin and 2.4. Extraction optimisation phenolic extraction could be one of the sources of high variation between reported values in a same plant material. 2.4.1. Selection of the solid-to-liquid ratio Furthermore, anthocyanins can be present in different Solid–liquid ratios of 1:30; 1:60, 1:90, 1:120, 1:200 were stud- compartments in the plant tissue, including vacuoles or in vacuolar ied in order to investigate the amount of solvent need to maximise inclusions. Thus, depending on the specific tissue characteristics, the yield of anthocyanin and phenolic compounds extraction. the optimum conditions for extraction to maximise yields may Freeze-dried pulp powder (0.5 g) was extracted under agitation vary (Giusti & Jing, 2008). for 1 h with 75% aqueous acetone solution containing 2% formic Simplex lattice and central composite design are two multivar- acid in the different solid–liquid ratios chosen. Extraction media iate experimental designs that allow investigating the effect of a was filtered under vacuum, solids were discarded, and the filtrate mixture or a set of factors under one or more responses. The was taken to an exact volume (100 mL) with the extraction solvent. advantage of these techniques is that it allows the researcher to The number of extractions was restricted to only one to avoid loses obtain knowledge of the whole system behaviour inside of the during the filtration process that could jumble the results. The ranges, to evaluate simultaneously multiple variables, and to pre- solid–liquid ratio with highest extraction yield of anthocyanin dict results using statically valid models (Ferreira et al., 2007). and total phenolic compounds content was fixed for the next There is a wide range of applications of this statistical tool in the experimental step in the optimisation process. Three independent literature from bioprocesses in enzyme production (Ries & Alves replications were performed and analysed in triplicate. ANOVA Macedo, 2011) to extraction conditions for phenolics (Cacace & tests and means comparisons using Tukey test (p 6 0.05) were per- Mazza, 2002; Monrad, Srinivas, Howard, & King, 2012). formed in Statistica 7.0 (StatSoft, Inc.). V.C. Bochi et al. / Food Chemistry 164 (2014) 347–354 349
2.4.2. Evaluation of solvent effects by simplex lattice design (SLD) 2.05%, see Supplementary material) variables. Eleven extraction Using the best solid–liquid ratio determined in the previous conditions were tested which included 22 complete factorial, 4 experiment, extraction efficiency with different proportions of ace- axial points, and 3 repetitions of central conditions. Total mono- tone, water, and ethanol were studied aiming to determine the sol- meric anthocyanin content (TMA, Y1, mg/g freeze-dried sample) vent combination to achieve higher extraction yields with lower and total phenolic compounds content (TPC, Y2, mg/g freeze-dried percentage of organic solvents. A multivariate statistical technique sample) were the two responses studied evaluated as described in for optimisation of mixtures, the simplex centroid design with Sections 2.5 and 2.6. axial points, was chosen due to the possibility of study all solvent Results were used to build a regression model with linear, qua- proportions (0–100%) by using a reduced number of experiments. dratic, and interactive components (Eq. (2)). Moreover, this design allowed us to examine possible interaction effect among variables. 2 2 1 2 y ¼ b0 þ b1x1 þ b2x2 þ b11x1 þ b22x2 þ b12x x þ e ð2Þ All conditions tested are presented in Table 1. In this design, the different conditions tested form a triangle, with pure components All experiments in the SLD and RSM were performed in random in the vertex, representing 100% of one of each solvent. Middle order to minimise the effects of unexplained variability in the points in each side representing a binary mixture (1:1), the center observed responses due to systematic errors. ANOVA, regression point as a ternary mixture (1:1:1), and axial points representing results, and surfaces from fitted models were obtained using Stat- 2/3 of one of the solvents and 1/6 for the others. This design per- istica 7.0 (StatSoft). The final extraction procedure, using optimal conditions, was validated based on predictions of the developed mits the evaluation of linear (b1, b2, and b3), quadratic (b12, b13, model. and b23), and special cubic models (b123) (Eq. (1)) for the response under study.
Y ¼ b1X1 þ b2X2 þ b3X3 þ b12X1X2 þ b13X1X3 þb23X2X3 þ b123X1X2X3 2.4.4. Comparison among extraction conditions and extraction ð1Þ efficiency with different berry matrices Optimised conditions determined after the series of experi- Total monomeric anthocyanin content (TMA, Y1, mg/g freeze- ments were compared to another extraction methodology reported dried sample) and total phenolic compounds content (TPC, Y2, in the literature (Rodriguez-Saona & Wrolstad, 2001). Freeze-dried mg/g freeze-dried sample) were the two responses evaluated as fine powders were exhaustively extracted with a mixture of 70/30 described in Sections 2.5 and 2.6. The proportion of solvents that (v/v) of acetone in acidified water (0.01% HCl) under maceration. gave the highest response with the lowest concentration of organic The resulting filtrates were combined, shaken, placed in a separa- solvent was fixed for the following experiments of time and acid tory funnel with chloroform (1:2 acetone: chloroform v/v), and concentration optimisation. stored overnight at 5 °C. The aqueous portion (top portion) was collected and placed on a rotary evaporator at 38 ± 2 °C until all 2.4.3. Time and acid concentration optimisation using response surface residual acetone was evaporated. The aqueous extract was made methodology (RSM) up to a known volume with acidified distilled water (0.01% HCl). Time and acid concentration were the two last variables For the optimised procedure, extracts were placed in a rotatory optimised using the response surface methodology. The best evaporator to remove organic solvent and made up to a known vol- solid–liquid ratio and mixture of solvents to obtain the highest ume using diluted formic acid solution (0.35%, pH 2.52 ± 0.01). concentration of anthocyanin and phenolic compounds were fixed Freeze-dried powder of Ceylon gooseberry pulp samples and entire based on results obtained from the previous experiments. At these berries of blueberry (Vaccinium myrtillus), blackberry (Rubus ap., conditions, a five levels ( a, 1, 0, +1, +a) central composite rota- Guarani cultivar), Pitanga Brazilian cherry (Eugenia uniflora L.), tional design was used to optimise extraction time (X1, from 17.6 and grape (Vitis labrusca, Isabel cultivar) were used in this to 102.43 min) and formic acid concentration (X2, from 0.35% to experiment.
Table 1 Solvent optimisation experiments: simplex lattice design and observed responses.
Run number Independent variables Responses
a a Acetone (X1) Ethanol (X2) Water (X3) TPC TMA Coded value Real value (%) Coded value Real value (%) Coded value Real value (%) 1 1 100 0 0 0 0 5.53 0.72 3 0 0 1 100 0 0 8.88 2.60 5 0 0 0 0 1 100 9.53 2.84 7 1/2 50 ½ 50 0 0 11.02 2.76 11 1/2 50 0 0 1/2 50 10.91 2.64 9 0 0 ½ 50 1/2 50 10.02 3.00 13 2/3 66.7 1/6 16.7 1/6 16.7 10.8 2.88 15 1/6 16.7 2/3 66.7 1/6 16.7 14.32 2.79 16 1/6 16.7 1/6 16.7 2/3 66.7 12.98 3.18 17 1/3 33.3 1/3 33.3 1/3 33.3 13.3 3.11 2 1 100 0 0 0 0 5.31 0.25 4 0 0 1 100 0 0 7.8 1.95 6 0 0 0 0 1 100 9.82 2.70 8 1/2 50 1/2 50 0 0 10.23 2.29 12 1/2 50 0 0 1/2 50 10.27 2.61 10 0 0 1/2 50 1/2 50 10.4 2.88 14 1/3 33.3 1/3 33.3 1/3 33.3 10.73 2.87
a Results were expressed as mg of cyanidin-3-glucoside equivalent/g of freeze-dried sample for total monomeric anthocyanins (TMA) and mg of gallic acid equivalent/g of freeze-dried sample for total phenolic compounds (TPC). 350 V.C. Bochi et al. / Food Chemistry 164 (2014) 347–354
2.5. Total monomeric anthocyanin content (TMA) 3.1. Solid–liquid ratio effect
Monomeric anthocyanin content was determined by using the Considering that solvents will extract target compounds until pH-differential method described by Giusti and Wrolstad (2001). saturation or until reaching to equilibrium with the sample, the This method is based on the anthocyanin equilibrium forms ratio between sample solids and solvent is one of the major vari- obtained under acid (pH 1.0) and in a higher pH (pH 4.5) condi- ables that could affect the extraction efficiency. Aiming to ensure tions. The coloured flavylium cation forms in acid conditions are that all variables were efficiently studied, solid–liquid ratios were measured against a correction in basic conditions, when colourless evaluated separately and fixed using the best condition before hemiketal forms of anthocyanins are present. continuing with the optimisation experiments. Results of solvent Readings were performed at 520 and 700 nm and the concen- (75% aqueous acetone containing 2% formic acid) to solid (freeze- tration of total monomeric anthocyanin content was calculated dried powder) proportion experiments are illustrated in Fig. 1.A according to the Eq. (3). Results were expressed as mg of cyani- significant increase in phytochemical extraction yields was din-3-glucoside equivalents (CGE) per gram of freeze-dried pulp observed when the method was performed using 1:90 solid–liquid and, using values obtained for water loss after freeze-drying pro- ratio for total anthocyanin content response, and 1:120 for total cess, results were converted to mg per 100 grams of pulp fresh phenolic compounds. Thus, solid–liquid ratio was fixed in 1:120 weight. Values of 449.2 and 26.900 were considered as molecular for the next experiments, since no changes were observed with weight (MW) and molar absorptivity (e) of cyanidin-3-glucoside, proportions higher than it and this value was suitable for both respectively (Eq. (4)). responses.
A ¼ðA A Þ ðA A Þ ð3Þ 520nm 700nm pH1:0 520nm 700nm pH4:5 3.2. Solvent effects
Monomeric anthocyanin pigment ðmg=litreÞ Acidified methanol, ethanol, and acetone are often used as ¼ðA MW DF 1000Þ=ðe 1Þð4Þ organic solvents for anthocyanin extraction (Giusti & Jing, 2008). Nevertheless, extraction with organic solvents has economic and MW: molecular weight; DF: dilution factor. Since it is a environmental disadvantages, and the concept of ‘‘green chemis- response used for evaluation of extraction variables in the optimi- try’’ encourages the development and the use of less hazardous sation experiments, quality assurance was checked using the rela- processes and materials without reduction in efficiency (Chemat, tive standard variation (RSD) for results obtained from extractions Vian, & Cravotto, 2012). Accordingly, solvent extraction of anthocy- (n = 3) over 3 days of analysis. anin should be optimised for the maximum response using less organic solvents. Thus, water was chosen to be studied in combina- 2.6. Total phenolic compounds content (TPC) tion with ethanol and acetone since native chemical structures of anthocyanins are usually linked to sugar moieties having increased Total phenolic content was measured using the Folin–Ciocalteu water solubility (Giusti & Jing, 2008). Ethanol was chosen instead method as described by Singleton and Orthofer (1999) and modi- of methanol as extraction solvent due to the high human toxicity fied by Scherer and Godoy (2014). This method is based on the of methanol (EPA, 2013). Moreover, low ethanol concentrations reaction between phenolic compounds and sodium carbonate to could be used in some food industries processes, as it has already form phenolates, which are able to reduce the reagent mixture been classified as GRAS (Generally Recognized as Safe) by the Folin–Ciocalteu. Color in the reaction mixture changes from yellow FDA (Food and Drug Administration, USA) for some food additive to blue and is monitored at 760 nm. The extraction solvent was preparations (FDA, 2012). used for the blank. In the experiment with different solvent (mix- Results from complete SLD obtained in the solvent optimisation ture design), one curve was performed for each mixture of sol- experiments are shown in Table 1. The lowest values for anthocy- vents. All results were presented as mg of gallic acid equivalents anin and total phenolic content were obtained in the experiment (GAE) per gram of freeze-dried sample and, using values obtained number 1 which corresponds to extraction with pure acetone. for water loss after freeze-drying process, results were converted Extraction with pure ethanol (experiment 3) and pure water (exper- to mg per 100 grams of pulp fresh weight. Quality assurance was iment 5) showed better extraction efficiency than with pure ace- checked using results obtained for the calibration curve over differ- tone (experiment 1). However, the best results were obtained ent days of analysis (n = 3).
3. Results and discussion
The total content of monomeric anthocyanins and phenolic compounds were the main responses used during optimisation studies and the relative standard deviation (RSD) values were checked for quality assurance. Thus, for the average value of 3.01 ± 0.22 mg of cyaniding-3-glucoside equivalents/g of freeze dried sample, RSD values in different days of analysis ranged from 3.2% to 4.3%. For the total content of phenolic compounds, a linear range was obtained from10 to 65 lg/ml and the regression coeffi- cient (r2) was equal to 0.9993 ± 0.0008, using a seven-point gallic acid calibration curve. Relative standard deviation (RSD) at the lowest, the intermediate, and the highest concentration tested for calibration curves were 4.66%, 3.67%, and 2.83%, respectively. Sec- Fig. 1. Total monomeric anthocyanin (TMA) and total phenolic content (TPC) of Ceylon gooseberry freeze-dried sample extracted with different solid/liquid ratio w/v. ond Horwitz & Albert, 1995, all values observed for RSD in both Extraction conditions: 1 h under agitation with 2% of formic acid in 75% acetone in methods are acceptable and indicate satisfactory method water. Results are mean ± standard deviation. Different letters in the same group precision. represent significant differences (Tukey test; p 6 0.05). V.C. Bochi et al. / Food Chemistry 164 (2014) 347–354 351
Table 2 with the highest desirability revealed that the best proportion of * ANOVA results of regression models estimated from SLD for solvent optimisation and solvents was 20% of acetone with 80% of water. The same propor- fitness of quadratic model. tion was obtained with all coefficients in the model, but adjusting ANOVA for the highest desirability in a proportion with the lowest amount 2 Source SS DF F-value p-Value R of organic solvents (ethanol and acetone). TMAa Aiming to validate the obtained fitted model for anthocyanin Linear 5.37 2 7.44 0.006 0.52 content new experiments were performed with different propor- Quadratic 4.15 3 16.74 0.0002 0.91 tions between solvents and in the best condition achieved (Table 3). Cubic 0.003 1 0.03 0.88 0.91 Three independent experiments were performed in each condi- TPCb tion and analysed. All results obtained for anthocyanin content Linear 10.7 2 0.94 0.41 0.12 Quadratic 50.9 3 6.58 0.008 0.68 are in agreement with the model predictions and within the Cubic 0.6 1 0.23 0.64 0.69 confidence interval (95%). Notwithstanding the model obtained Fitness of purpose for total monomeric anthocyanin’s quadratic model for total phenolic had significant lack of fitness (p-value of Source SS DF F-value p-value 0.00008) being not useful for predictions, solvent mixtures used TMAa for the anthocyanin model validation were also monitored for this Model 9.52 5 23.05 0.000017 response. Results in conditions B and D (Table 3) were inside of Total Error 0.91 11 the interval of confidence. However, as expected for a non-fitted Lack of fit 0.46 4 1.79 0.24 model, conditions A and C were higher than predicted values. Pure error 0.45 7 Condition D showed the highest content of anthocyanin and b TPC total phenolic compounds which is in agreement with the Model 61.56 5 4.78 0.01 Total error 28.32 11 optimum condition predicted by the anthocyanin quadratic Lack of fit 27.08 4 38.19 0.00008 model. Pure error 1.24 7 Fig. 2 shows the surface obtained with the valid model for total * Simplex lattice design. monomeric anthocyanin content response, only with significant a Total monomeric anthocyanins. coefficients (Y =+0.533⁄ X1 + 2.456 ⁄ X2 + 2.911 ⁄ X3 +4.780⁄ X1 b Total phenolic compounds. ⁄ X2 + 4.110 ⁄ X1 ⁄ X3 + 0.100). Even though the best condition was the mixture of 20% acetone, 80% water, it is possible to deter- with a mixture of solvents (experiments 15 and 16). Thus, for total mine that extraction with pure water will also produce acceptable phenolic content the highest yield (14.32 mg of GAE/g freeze-dried results. In agreement with this observation, some recent works (Li, sample, experiment 15) was obtained with ethanol as the main sol- Fabiano-Tixier, Vian, & Chemat, 2013; Li et al., 2012; Petersson, Liu, vent in the mixture composed of 1/6, 2/3, and 1/6 fractions of ace- Sjöberg, Danielsson, & Turner, 2010; Puértolas, Cregenzán, Luengo, tone, ethanol, and water in the whole solvent volume, respectively. Álvarez, & Raso, 2012) have proposed water as environmental- A similar behaviour was observed for total monomeric anthocya- friendly solvent for extraction of phenolic compounds and antho- nin where the highest concentration was obtained in experiment cyanins, in some cases with similar yield to organic solvents com- 16, a mixture of all solvents with higher amount of water than ace- monly used. However, in previous experiments, pure water was tone and ethanol (1/6, 1/6, and 2/3 fractions of acetone, ethanol, applied in combination with pressure and high temperatures and water in the whole solvent mixture, respectively). (Petersson et al., 2010), micro-wave-assisted techniques (Li et al., ANOVA from regression models and fitness of purpose for qua- 2012, 2013), and pulsed-eletric-field-assisted (Puértolas et al., dratic models are shown in Table 2. Significant quadratic models 2012) extraction with results that were similar to values reported were obtained for both responses; however, just the model when using organic solvents. In our experiments, we obtained high obtained for anthocyanin content has no lack of fit and could be extraction yields without assistance of additional extraction tech- used for make predictions. Significant coefficients were obtained nologies other than simple homogenisation. for all variables studied (0.55 ± 0.2 for b1, 2.29 ± 0.2 for b2, and 6 2.75 ± 0.2 for b3 with p-values of 0.018, 0.1 10 , and 3.3. Time and acid concentration effect 0.1 10 6, respectively) and almost all interactions showed an important synergistic effect between acetone and ethanol For the studies of time and acid effects the extracting solvent
(4.82 ± 0.91 for b12), and between acetone and water (4.15 ± 0.91 was fixed according to the results of previous experiments at 20% for b13). However, the interaction coefficient between ethanol of acetone in water as the optimum for anthocyanin extraction. and water (b23) was no significant for the studied response. Using this proportion, time and acid concentration were optimised Because of it, this coefficient was removed from the model before in a central composite design. There was almost no difference in the predictions (see estimated regression coefficients, standard error anthocyanin and total phenolic content obtained from all different and p-values at Supplementary material). The model prediction experiments (see complete experimental design and results in
Table 3 Validation results for quadratic model obtained from SCD* in solvent optimisation experiments.
Condition Acetone Ethanol Water TMAa (mg/g freeze-dried sample) TPCb (mg/g freeze-dried sample) Predicted Values Interval of Experimental Values Predicted Values Interval of Experimental Values confidence confidence
X1 X2 X3 95% +95% -95% +95% A 0.37 0 0.63 2.99 2.59 3.4 2.73 + 0.02 11.89 9.82 13.96 13.08 + 0.48 B 0.25 0.32 0.43 2.99 2.77 3.21 3.07 + 0.04 11.6 10.46 12.73 13.28 + 0.13 C 0.25 0.75 0 2.87 2.52 3.22 2.78 + 0.02 11.34 9.53 13.15 12.84 + 0.14 D 0.2 0 0.8 3.1 2.76 3.42 3.19 + 0.05 11.64 9.92 13.35 14.21 + 0.06
* Simplex lattice design. a Total monomeric anthocyanins expressed as cyanidin-3-glucoside equivalent. b Total phenolic compounds expressed as gallic acid equivalent. 352 V.C. Bochi et al. / Food Chemistry 164 (2014) 347–354
3.4. Comparison between extraction conditions method efficiency with different berry matrices
There is a diversity of extraction methods for anthocyanins and phenolic compounds which are recognised as able to be applied from berries to more complex samples. Rodriguez-Saona and Wrolstad (2001) have published a rather universal protocol that allows for the extraction of anthocyanin from a range of different
materials (Duan, Wu, Strik, & Zhao, 2011; Lima, Melo, Pinheiro, & Guerra, 2011) with a subsequent partition in chloroform that par- tially purifies the pigments. Extraction is exhaustively performed with 70% of acetone in water with hydrochloric acid (0.01%). This procedure was reproduced using freeze-dried powder of Ceylon gooseberry and compared to the extraction in the optimised and initial conditions aiming to evaluate the efficiency of final extrac- tion method. Moreover, it was evaluated the applicability of the optimised extraction in other berry sample matrices (entire grapes, Pitanga cherry, blackberry, and blueberry). All results obtained with our optimised protocol were compared to the results obtained with exhaustive extraction described above (Table 4). For Ceylon gooseberry samples, optimised conditions showed higher (p 6 0.05) values for TMA (Table 4) and TPC (13.96 ± 0.06
Fig. 2. Total monomeric anthocyanin (TMA) content surface obtained with mg GAE/g of freeze-dried sample) when compared to the initial quadratic model for solvent mixture optimisation using simplex centroid design procedure (12.57 ± 0.05 mg GAE per gram of freeze-dried pulp (SCD). Results are expressed as mg of cyanidin-3-glucoside/g of freeze-dried sample, respectively) representing an increase of 31.54% and sample. 11.06%, respectively. RSD using our optimised conditions was lower than the values obtained with the traditional extraction Supplementary material). ANOVA from regression analyses showed conditions showing an improvement in method repeatability. that linear and quadratic models are not significant to explain the Moreover, for all matrices tested, ANOVA revealed no signifi- oscillation among the results. In other words, in the range studied, cant differences between our optimised conditions determined in acid and time had no effect on the total phenolic compounds and this work and the exhaustive traditional extraction procedure. total monomeric anthocyanin content yields obtained. Results ran- Thus, our optimised extraction conditions were applied to the ged from 3.32 to 3.65 mg of CGE/g and from 12.55 to 14.21 mg of extraction of phenolics and anthocyanins from Ceylon gooseberry GAE/g of freeze-dried pulp sample for total monomeric anthocya- pulp samples and in all tested matrices with similar extraction nin content and total polyphenol content, respectively. yields than those obtained with the exhaustive extraction proce- Considering that short extraction times are desirable and lower dure with 70% acetone. For sources other than berry fruits and with acid concentration are desirable to avoid hydrolysis and degrada- a different cellular structure, the optimum extraction conditions tion of target compounds, 0.35% of formic acid and 17.6 min were for maximised compound recovery should be checked prior to the conditions chose as the optimum for extraction of anthocyanins their chemical analyses. Both extraction methods showed satisfac- and phenolic compounds of Ceylon gooseberry. Larger ranges could tory repeatability, with RSD (from 1.9% to 5.4%, Table 4) inside of be tested for extraction time and acid concentration aiming to find a the limits described by Horwitz and Albert (1995) in the range of valid model. Nevertheless, the obtained results probably are very concentrations found in fruit matrices. close to the real amount in the pulp and extreme conditions could ANOVA showed significant differences among mean values of result in less stable extracts. total anthocyanin content determined in berry samples. Ceylon
Table 4 Extraction comparison results for total monomeric anthocyanin content in Ceylon gooseberry freeze-dried pulp and other berry fruit matrices.
Sample Extraction condition TMA content (mg/g freeze-dried sample) RSD (%) Ceylon gooseberry pulp (Dovyalis hebecarpa) A 2.79 ± 0.19 6.73 B 3.67 ± 0.09 3.67 C 3.29 ± 0.22 3.29 Mean value⁄ 3.48b Entire blueberry (Vaccinium myrtillus) B 5.25 ± 0.39 5.24 C 5.32 ± 0.27 5.29 Mean value 5.28a Entire blackberry (Rubus ap., Guarani cultivar) B 5.75 ± 0.70 5.36 C 4.66 ± 1.21 4.78 Mean value 5.20a Entire Pitanga Brazilian cherry (Eugenia uniflora L.) B 2.40 ± 0.31 2.49 C 1.91 ± 0.17 1.90 Mean value 2.15c Entire grape (Vitis labrusca, Isabel cultivar) B 2.66 ± 0.25 2.56 C 2.32 ± 0.19 2.71 Mean value 2.49c
RSD: relative standard deviation. Results are mean values (n = 3) plus standard deviations. Same lower case letter in the column means no significant differences at 95% of confidence (p 6 0.05) by Tukey test. A: Extraction at initial conditions which were 2% of formic acid in 75% of acetone in water under one hour of mixing. B: Final optimised conditions which were 0.35% of formic acid in 20% of acetone in water under 18 min of mixing. C: Exhaustive extraction with 0.01% of HCl in 70% of acetone in water as reported by Rodriguez-Saona and Wrolstad (2001). ⁄ Mean value of results obtained in conditions B and C. V.C. Bochi et al. / Food Chemistry 164 (2014) 347–354 353 gooseberry pulp sample showed higher anthocyanin levels than Chemat, F., Vian, M. A., & Cravotto, G. (2012). Green extraction of natural products: entire Pitanga cherry and grape berries (Isabel cultivar) and lower Concept and principles. International Journal of Molecular Sciences, 13(7), 8615–8627. than blueberries and blackberries. Since, Ceylon gooseberry peels Chiva-Blanch, G., & Visioli, F. (2012). Polyphenols and health: Moving beyond have a velvety texture and it is usually removed before consump- antioxidants. Journal of Berry Research, 2, 63–71. tion, only pulp fruit part was considered as edible part and used De Rosso, V. V., & Mercadante, A. (2007). HPLC–PDA–MS/MS of anthocyanins and carotenoids from Dovyalis and Tamarillo fruits. Journal of Agricultural and Food for analysis. As previously reported (Cevallos-Casals, Byrne, Okie, Chemistry, 55, 9135–9141. & Cisneros-Zevallos, 2006; Lee & Wrolstad, 2006), higher concen- Dobson, P., Grahan, J., Stewart, D., & Brennan Hackett, C. A. (2012). Over-seasons tration of polyphenols are normally found in the exocarp than in analysis of quantitative trait loci affecting phenolic content and antioxidant capacity in raspberry. Journal of Agriculture and Food Chemistry, 60, 5360–5366. the mesocarp fruit part. Thus, results for whole Ceylon gooseberry Duan, J., Wu, R., Strik, B. C., & Zhao, Y. (2011). Effect of edible coatings on the quality fruits are expected to be higher and possibly closer to blackberry of fresh blueberries (Duke and Elliott) under commercial storage conditions. and blueberry anthocyanin levels. Finally, Ceylon gooseberry pulp Postharvest Biology and Technology, 59(1), 71–79. EPA (United States Environment Protection Agency) (2013). Toxicological review of can be considered a source of anthocyanin and other phenolic com- methanol (noncancer), (CAS No. 67-56-1). National Service Center for pounds with concentration values converted to fresh weight Environmental Publications (NSCEP). EPA/635/R-11/001Fa. 212p. (49.04 ± 0.08 mg CGE/100 g and 187.06 ± 0.08 mg of GAE/100 g of FDA (United States Food and Drug Administration) (2012). Draft guidance for FW) similar to other gooseberry species (2.4–43.3 mg of CGE/ industry: Assessing the effects of significant manufacturing process changes, including emerging technologies, on the safety and regulatory status of food 100 g FW) and raspberries (1.3–88.4 mg of cyanidin-3-gluoside ingredients and food contact substances, including food ingredients that are equivalents/100 g FW 129–184.6 mg GAE/100 g) (Dobson, Grahan, color additives. U.S. Department of Health and Human Services, Food and Drug Stewart, & Brennan Hackett, 2012; Pantelidis, Vasilakakis, Administration, Center for Food Safety and Applied Nutrition. Available at: http://www.fda.gov/Food/GuidanceRegulation/ Manganaris, & Diamantidis, 2007). GuidanceDocumentsRegulatoryInformation/ IngredientsAdditivesGRASPackaging/ucm300661.htm. Ferreira, S. L. C., Bruns, R. E., da Silva, E. G. P., dos Santos, W. N. L., Quintella, C. M., 4. Conclusion David, J. M., de Andrade, J. B., Breitkreitz, C. M., Jardim, I. C. S. F., & Barros Neto, B. (2007). Statistical designs and response surface techniques for the optimization Extraction conditions were optimised and maximum anthocya- of chromatographic systems. Journal of Chromatography A, 1158(1), 2–14. Giusti, M. M., & Jing, P. (2008). Analysis of Anthocyanins. In C. Socaciu (Ed.), Food nin yields were obtained in experiments with high proportions of colorants: Chemical and functional properties. 978-0-8493-9357-0. Taylor and water (from 50% to pure water, with the optimum conditions using Francis Group, CRC Press. Chapter 6.3. 80% of water with 20% acetone). Since pH values in final solvent Giusti, M. M., & Wrolstad, R. E. (2001). Characterization and measurement of anthocyanins by UV-visible spectroscopy. Current Protocols in Food Analytical media did not reach to values above 2.0, acid concentration had Chemistry, F1.2.1–F1.2.13. no effect in anthocyanin and phenolic extraction and neither did Horwitz, W., & Albert, R. (1995). Precision in analytical measurements: Expected the extraction time in the ranges tested. Consequently, both vari- values and consequences in geochemical analyses. Fresenius’ Journal of Analytical Chemistry, 351, 507–513. ables were fixed in the minimum values studied. The optimised Lattanzio, V., Kroon, P. A., Quideau, S., & Treutter, D. (2009). Plant phenolics – condition, using less organic solvent in a shorter time, produced secondary metabolites with diverse functions. In F. Daayf & V. Lattanzio (Eds.), phytochemical yields comparable to those obtained with other tra- Recent advances in polyphenol research (Vol. 1). http://dx.doi.org/10.1002/ ditional methods using harsh chemicals. Moreover, Ceylon goose- 9781444302400. Chapter 1. Lee, J., & Wrolstad, R. E. (2006). Extraction of anthocyanins and polyphenolics from berry is a source of anthocyanin and other phytochemical with blueberry processing waste. Journal of Food Science, 69, 564–573. http:// great potential to be used with industrial purposes, since it can dx.doi.org/10.1111/j.1365-2621.2004.tb13651.x. be extracted with pure water. Li, Y., Fabiano-Tixier, A. S., Vian, M. A., & Chemat, F. (2013). Solvent-free microwave extraction of bioactive compounds provides a tool for green analytical chemistry. TrAC, Trends in Analytical Chemistry. http://dx.doi.org/10.1016/ Acknowledgements j.bbr.2011.03.031. Li, Y., Han, L., Ma, R., Xu, X., Zhao, C., Wang, Z., & Hu, X. (2012). Effect of energy density and citric acid concentration on anthocyanins yield and solution The authors are thankful to Conselho Nacional de Desenvolvi- temperature of grape peel in microwave-assisted extraction process. Journal of mento Científico e Tecnológico (CNPq, Brazil) for a doctorate fellow- Food Engineering, 109(2), 274–280. ship, to the Coordenação de Aperfeiçoamento de Pessoal de Nível Lima, V. L. A. G. D., Melo, E. A., Pinheiro, I. O., & Guerra, N. B. (2011). Antioxidant capacity of anthocyanins from acerola genotypes. Ciência e Tecnologia de Superior (Capes, Brazil), Fulbright Commission (USA), and The Ohio Alimentos, 31(1), 86–92. State University (OSU, Columbus, OH, USA) for the support during Mertz, C., Cheynier, V., Günata, Z., & Brat, P. (2007). Analysis of phenolic compounds the study period abroad. We are also grateful to Josenilda Palmeira in two blackberry species (Rubus glaucus and Rubus adenotrichus) by high- performance liquid chromatography with diode array detection and Rodriguez for supplying the sample used in this work. electrospray ion trap mass spectrometry. Journal of Agricultural and Food Chemistry, 55(21), 8616–8624. Naczk, M., & Shahidi, F. (2006). Phenolics in cereals, fruits and vegetables: Appendix A. Supplementary data occurrence, extraction and analysis. Journal of Pharmaceutical and Biomedical Analysis, 41(5), 1523–1542. http://dx.doi.org/10.1016/j.jpba.2006.04.002. Supplementary data associated with this article can be found, in Monrad, J. K., Srinivas, K., Howard, L. R., & King, J. W. (2012). Design and optimization of a semi continuous hot–cold extraction of polyphenols from the online version, at http://dx.doi.org/10.1016/j.foodchem.2014. grape pomace. Journal of Agricultural and Food Chemistry, 60(22), 5571–5582. 05.031. Morton. J. F., Dowling. C. F. (1987). Fruits of warn climates. ISBN: 0-9610184-1-0.
Rodriguez-Saona, L. E., & Wrolstad, R. E. (2001). Extraction, isolation, and Shahidi, F., & Naczk, M. (2003). Biosynthesis, classification and nomenclature of purification of anthocyanins. Current Protocols in Food Analytical Chemistry, phenolics and nutraceuticals. In M. Naczk & F. Shahidi (Eds.), Phenolics in food F1.1.1–F1.1.11. and nutraceuticals. 0-203-59485-1. CRC Press. Chapter 1. Scherer, R., & Godoy, H. (2014). Effects of extraction methods of phenolic Singleton, V. L., Orthofer, R., & Lamuela-Raventos, R. M. (1999). Analysis of total compounds from Xanthium strumarium L. and their antioxidant activity. phenols and others oxidation substrates and antioxidants by means of Folin- Revista Brasileira de Plantas Medicinais, 16(1), 41–46. http://dx.doi.org/10.1590/ Ciocaulteau reagent. Methods in Enzymology, 299, 152–172. S1516-05722014000100006. Zafra-Stone, S., Yasmin, T., Bagchi, M., Chatterjee, A., Vinson, J. A., & Bagchi, D. Shahidi, F., & Naczk, M. (2003). Methods of analysis and quantification of phenolic (2007). Berry anthocyanins as novel antioxidants in human health and disease compounds. Phenolics in food and nutraceuticals. CRC Press. prevention. Molecular Nutrition & Food Research, 51(6), 675–683.